use openblas instead of gslcblas

[smuzaffar]

• Build OpenBlas with DYNAMIC_ARCH=1
• Drop gslcblas library from gsl toolfile and add openblas dependency to make sure that any cmssw package using gsl links openblas instead of gslcblas
• set GSL_CBLAS_LIB so that gsl-config properly return -lopenblas instead of -lgslcblas
• Fixed geneva, thepeg and herwigpp specs to link against openblas

[smuzaffar]

please test

[cmsbuild]

The tests are being triggered in jenkins.
https://cmssdt.cern.ch/jenkins/job/ib-run-pr-tests/1253/console Started: 2019/06/28 17:40

[cmsbuild]

A new Pull Request was created by @smuzaffar (Malik Shahzad Muzaffar) for branch IB/CMSSW_11_0_X/gcc700.

@cmsbuild, @smuzaffar, @gudrutis, @mrodozov can you please review it and eventually sign? Thanks.
cms-bot commands are listed here

[cmsbuild]

-1

Tested at: a849fb5

You can see the results of the tests here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-fda669/1253/summary.html

I found follow errors while testing this PR

Failed tests: UnitTests

• Unit Tests:

I found errors in the following unit tests:

---> test DetectorDescriptionRegressionTestDOMCount had ERRORS

[cmsbuild]

Comparison job queued.

[cmsbuild]

Comparison is ready
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-fda669/1253/summary.html

Comparison Summary:

• No significant changes to the logs found
• Reco comparison results: 5 differences found in the comparisons
• DQMHistoTests: Total files compared: 33
• DQMHistoTests: Total histograms compared: 3255670
• DQMHistoTests: Total failures: 2
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 3255334
• DQMHistoTests: Total skipped: 334
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 32 files compared)
• Checked 137 log files, 14 edm output root files, 33 DQM output files

[slava77]

There is one small difference in a jet b-tag discriminant variable

WF # | Differences
-- | --
all_OldVSNew_RunJetHT2016EreMINIAODwf136p7611 | 1

It's hard to say which one (I only plot by the vector index); it may even be one of the mxnet-based cases.
@hqucms we may need some further investigation.

@smuzaffar please clarify if this change (alone, or at least without much else related) can be made available in the DEVEL branch IBs.
Thank you.

[smuzaffar]

@slava77 , DEVEL IBs are broken due to tensorflow update. I can include it in gcc8 or root 618 IBs

[hqucms]

There is one small difference in a jet b-tag discriminant variable

WF # | Differences
-- | --
all_OldVSNew_RunJetHT2016EreMINIAODwf136p7611 | 1

It's hard to say which one (I only plot by the vector index); it may even be one of the mxnet-based cases.
@hqucms we may need some further investigation.

@slava77 Is this difference in AK4 jets or AK8 jets? Only AK8 jets have MXNet-based taggers.

[slava77]

@slava77 Is this difference in AK4 jets or AK8 jets? Only AK8 jets have MXNet-based taggers.

in AK8

[hqucms]

@slava77 Is this difference in AK4 jets or AK8 jets? Only AK8 jets have MXNet-based taggers.

in AK8

OK I will take a look.

[hqucms]

@slava77 I confirm that there are some small numerical changes on the MXNet-based tagger outputs after changing the BLAS library. The differences are small enough so they should not have any impacts on the physics, but will unfortunately affect the reproducibility of existing NanoAOD workflows.

[slava77]

@hqucms
thank you for checking.

This sounds "good enough" for 11_X then.
Perhaps if all is good after a relval cycle we could consider a backport.

[smuzaffar]

+externals
@fabiocos , would you like to get it in 11.0.0.pre3?

[cmsbuild]

This pull request is fully signed and it will be integrated in one of the next IB/CMSSW_11_0_X/gcc700 IBs (but tests are reportedly failing). This pull request will now be reviewed by the release team before it's merged. @davidlange6, @slava77, @smuzaffar, @fabiocos (and backports should be raised in the release meeting by the corresponding L2)

[fabiocos]

@smuzaffar @slava77 @hqucms how urgent is this/ i have already started the build, of course we can integrate, launch an IB and restart in parallel, if this is felt as urgent

[slava77]

@smuzaffar @slava77 @hqucms how urgent is this/ i have already started the build, of course we can integrate, launch an IB and restart in parallel, if this is felt as urgent

In the context of potential backports and the next pre potentially falling into August and slow response from validators (in August), it would be nice to have it in pre3.

[fabiocos]

+1

@smuzaffar I would suggest to start a build with package integrated, and restart the 11_0_0_pre3 build with it

[smuzaffar]

sure @fabiocos , I am triggering an IB now.

[fabiocos]

@smuzaffar thanks